DAOs, Dunbar's Number and Network Effects
spannocchi.eth
0xaeE6
May 2nd, 2022

… and what governance should be aware about

DAOs, short for Decentralized Autonomous Organizations, are a novel way to organize human enterprise with more freedom for members and fewer organizational barriers that allows a fluid coming and going of contributors and maximum focus on community and value creation.

Most DAOs start out small and attract members and contributors as they tackle challenges and solve problems. The speed of growth is governed by network effects, as this article will explain.

Network effects are weak in the beginning but can become formidable forces for growth - and for decline - once a certain threshold is reached and new users come - or leave - a DAO simply because others they know are doing the same thing.

Make sure network effects work in your favour.
Make sure network effects work in your favour.

We will describe in detail what network effects are and give hints about how they apply to DAOs and then focus on practical implications. While more traditional organizations like companies or NGOs can also be viewed as networks, DAOs allow members to join or leave with almost no friction. A solid understanding of the dynamics of growth and decline is invaluable to founders, pod leaders and contributors alike. We hope this article can do its part to help.

Takeaways

  • Network effects govern the growth and decline of DAOs
  • Governance should be aware of network saturation and culture at all times
  • Dunbar’s number means size matters, and larger groups splinter into fractions
  • Only a shared purpose or calling can maintain group cohesion as DAOs scale

Introduction to network effects

Theodore N. Vail, president of the American Telephone & Telegraph Company, was the first to mention the term network effects in Bell Corporation's 1908 annual report. He used it to argue that Bell had a natural monopoly and that consumers would ultimately benefit from the consolidation of telephone exchanges in the hands of his company. Fast-forward to the 1980s when Ethernet inventor Robert Metcalfe latched onto this idea to describe how a communications network’s value is proportional to the number of its users squared.

A network of two users has only one possible connection, while a network of three has three, one of four already has six, and so on. The number of connections, and hence the utility of the network scale is proportional to the number of nodes squared. Throughout this article, we will use nodes and users interchangeably, although, in reality, a node could represent many users, or one user could operate multiple nodes.

Illustrations of network effects, source: applico.inc
Illustrations of network effects, source: applico.inc

Social networks like Facebook are the best examples of network effects. With each new user, the likelihood of finding known friends on the network increases, as does the cost of switching to another network with fewer users. Markets become winner-take-all, if the competing networks are sticky, which means:

  • Utility derived by users from network effects exceeds utility from switching
  • Users have high costs being active on more platforms
  • Users have high switching costs

Users can easily be active on multiple DAOs, which dampens network effects. Within similar types of DAOs (e.g., DeFi DAOs like Sushi) switching is not expensive, except if a user already developed strong social ties to members of one DAO that aren’t members of the other DAO.

Crypto’s focus on interoperability has the effect of lowering switching costs intentionally, so networks are less sticky and more replaceable.

Demand-side economies of scale and critical mass

Economies of scale usually describe how the price of producing a good is reduced with the scale of production. One of a million laptops produced is much cheaper than if a gifted hardware engineer would make one-offs by hand. When cost is reduced, more people can afford a product, which increases demand, which is why this is called supply-side economies of scale.

Network effects, on the other hand, are demand-side economies of scale. As more and more users flock to a network, its value increases, and more users are willing to afford a higher price.

In other words, the willingness to pay, for a buyer, increases as the number of buyers or sellers for the business grows,” says Harvard Business School Professor Bharat Anand.

The point where the utility to a new user is equal to the price paid is called critical mass. From this point on strong network, effects can be observed and the network can often grow without advertising, just by word of mouth and momentum. “Once you’ve gained significant market share, you can often sit back and let the network effect take over,” Anand says, “Your existing buyers and sellers are, in effect, your sales force in attracting more buyers. You often have to do very little.

Mark Zuckerberg, CEO of Facebook, had an intuitive grasp of these dynamics when starting out. He reduced his initial market size by only allowing users with a Harvard email address to sign up, giving the product an exclusive and elite appeal on top. After Harvard’s market was saturated, he opened the product to Yale and Stanford students. The product was and still is free for users, so critical mass is reached quicker than with paid products.

Mike Novogratz, CEO of Galaxy Digital, draws a similar comparison: “One of the powerful things for the Ethereum narrative is valuing the Ethereum network kind of like we do Facebook the more network effects you get… and so you look at DeFi (decentralized finance) and stablecoins… and just that alone brings more prominence and network effects to Ethereum”.

For DAO founders the important lesson is to reduce the addressable market in the beginning and to focus on a specific niche so that critical mass can be reached. If an audience is too small, growth is limited from the get-go, if it’s too big it becomes hard and expensive to gain traction. The ideal audience for a DAO is big enough that it's reasonable to assume enough contributors and enough problems to solve can be found. But not so big that even an active 100 member DAO could easily go unnoticed.

Direct and indirect network effects

Social networks and instant messengers are prime examples of direct network effects. Each new user adds value to the network. Platforms like eBay, however, are different. Each new seller doesn’t add any value to sellers, if anything, the added competition increases price pressure. But the additional seller adds value for buyers, which now have a greater variety of goods to choose from or see lower prices. This attracts more buyers who add value to the sellers, attracting more sellers. These are indirect network effects.

Many DAOs who sell services are facing indirect network effects. New members add capacity, but without takers, this is meaningless. For DAO customers, added capacity means more services or quicker turnaround, which can lead to even more demand for new members. In a way, DAOs are ultimately scalable organizations because no formal hiring or firing process is involved. Contributors flock to DAOs where they can provide meaningful input and drift away when they are bored. They often undergo some sort of initiation process and perform small tasks to prove their mettle. Afterwards, they take on larger and larger tasks and get more and more rights and responsibilities, both from other members and coded into the contracts that govern the DAO’s inner workings.

Negative network effects – saturation, notification overload and outrage

We have only discussed positive network effects, where a virtuous circle of growth helps a network to attract more and more users, so far. But where there is light there must surely be shadow.

Network effects are just as effective when users leave. Metcalfe’s law states that the value of a network is proportional to the number of nodes squared. A 100 user DAO would have a proportional value of 10,000. If one user leaves, the proportional value drops to just 9,801, a two per cent loss associated with just a single user leaving. Because the network lost value, other users have fewer reasons to stick around and a cascade of loss can start.

We have seen this start to happen on Facebook and even in DAOs like Friends With Benefits. The toxic debate around Brantly.eth remaining a major delegate to ENS also threatened ENS-DAO with the onset of negative network effects.

The most common cause for negative network effects is saturation, the point where network capacity is exceeded. In the example of a telephone exchange, when as many users as the exchange can process, try to make a call, then the next one picking up the phone gets a busy signal. Bad user experience leads to frustrated users and diminished growth, or worse, users leaving.

In peer-to-peer systems capacity usually scales with users. BitTorrent is a classic example. The more people share files the more bandwidth the system has. But this is not the case for cryptocurrencies, where transactions must be verified by a majority of the miners or validators in a network. Ethereum users have intimate knowledge of what network saturation looks like, as most have seen transaction fees spike during NFT launches or coveted airdrops. Other cryptocurrencies picked up the slack and started the Layer One movement, most notably Solana and Terra. Networks like Arbitrum and Optimism built faster and cheaper layers on top of Ethereum, and are aptly called layer two networks.

Bitcoin and Ethereum scale via L2s, Avalanche scales via Subnets
Bitcoin and Ethereum scale via L2s, Avalanche scales via Subnets

For DAOs, saturation occurs in Discord chats, when too many people flood a timeline, and it becomes hard to follow a conversation. Constant outrage and a toxic communication culture can also lead to more sensitive users leaving. This changes the composition of the user base. As less aggressive users leave the ratio of moderate to toxic communicators becomes even worse, and ultimately dooms the DAO. A trickle of users leaving can lead to an exodus sooner rather than later.

Dunbar’s number and the power of myths

Growth is beautiful and bragging about five-digit Discord membership is fun, but what do so many people actually do in a DAO. Chats get flooded with messages, making it impossible to catch up or have a decent conversation. Identifying what to work on, and what tasks are available can become so hard that new members resign and become passive first, and leave a while later.

DAO governance needs to know how to deploy new members so that growth translates into traction and new members become valuable contributors. Two main points to watch out for are the size of groups and the clarity of purpose. It turns out that the maximum size for a group, where it still can easily maintain a feeling of belonging and cohesiveness was well established by Robert Dunbar.

Robert Dunbar is a British anthropologist who studied group sizes in primates. He found out that humans cannot maintain more than 150 meaningful relationships on average and the primary mechanism for group cohesion is gossip.

Illustration of Dunbar’s Number. Source: Wikimedia
Illustration of Dunbar’s Number. Source: Wikimedia

Dunbar’s number is an important breaking point in DAOs. As the number of members rises, a feeling of cohesion and “being in it together” evaporates and splinter groups form. These splinter groups often define themselves as different and better than other subgroups of the same organization, leading to infighting, deteriorating communications, loss of productivity and dissatisfaction.

One way to maintain a common purpose and group cohesion in larger groups is by defining a strong sense of shared identity with a clear and reiterated purpose, or even collective mythology. Christianity and the Catholic church could briefly unite a fractionalized medieval Europe across all boundaries to fight the “infidels” in the crusades.

Peter the Hermit preaching the First Crusade. Source: public domain
Peter the Hermit preaching the First Crusade. Source: public domain

In enlightened, modern society mythology is often replaced with good storytelling to define a common purpose and rally the members of organizations way past 150 members. DAO governance is well advised to keep their finger on the pulse of group dynamics and split large organizations into Pods and SubDAOs to keep members engaged.

Common values and purposes can hardly be reiterated too often to maintain a sense of belonging. DAO members working on community engagement often reach a point where they feel they repeated the same message ad nauseam, when in fact the audience was changing each time. It is important for governance to reiterate the necessity of keeping the spirit of the DAO alive and support community engagement

Practical implications of network effects for DAO governance

We have touched upon ways that DAOs can use network effects briefly in the previous sections and we will use this chapter to expand on that. Specifically what governance teams need to be aware of to successfully steer a decentralized organization.

Growing a DAO

When starting a DAO, the focus of governance is on growth. When the DAO isn’t born out of a successful product, like an NFT launch (BAYC and Loot come to mind), DAOs start with a few members who want to build something valuable and expect to solve a specific problem.

Reaching critical mass and finding an army of members who promote the DAO is crucial at this point. Rarely do treasuries allow generous ad spends or press budgets. All successful DAOs start with a strong wish to solve a specific problem or to form a community they want to build. Critical mass is reached earlier inside a smaller market because a significant market share requires less effort. Another route is to tackle a problem that attracts the attention of many. ConsitutionDAO is a perfect example of the latter. Buying a physical copy of the constitution was such a remarkable feat, that many jumped on board and the DAO could scale very quickly. But these are exceptions.

For a more commonplace example imagine a DAO that brews beer, let’s call it beerDAO. Instead of wanting to compete with international brands like Anheuser Busch, beerDAO would do much better to focus on creating content for homebrewing beer and selling starting kits or making a world-class local craft beer if and instantly drinkable product is desired. Small markets allow a DAO to reach a critical mass of members in a shorter time frame.

If members do not keep referring new members, governance must find out what is happening. If the answer cannot be obtained in conversations with members, deliberate experiments can at the very least make the DAO more interesting. Serendipitous breakthroughs are also much more likely if founders keep iterating.

Governance needs to watch out for the tell-tale signs of network effects:

  • Do existing members keep bringing new members?
  • Do members talk about being a part of the DAO without being prompted to do so?
  • Is there a lively ongoing dialogue in forums? What is the tone of the conversation?
  • Do friendships and connections form?

These dynamics cannot be forced but are hallmarks of a community forming. If members do not keep referring new members, governance must find out what is happening. Is the DAOs purpose not clear, or unattractive? Are the tasks not enticing enough, or are the rewards not satisfying? Is the DAOs culture not inviting? If the answer cannot be obtained in conversations with members, deliberate experiments can at the very least make the DAO more interesting. Serendipitous breakthroughs are also much more likely if founders keep iterating.

Keeping DAO members engaged

Governance wants to make sure new members grow to be engaged, valuable contributors, and stick around. If new members express boredom and unhappiness this will affect the overall dynamics in the DAO and could lead to users leaving and negative network effects set in.

Members who join, but don’t participate are not necessarily a bad thing. Sometimes new members want to get a lay of the land before they feel confident enough to sign up for tasks or participate in conversations. Lurkers represent a potential waiting for DAO governance to activate. Polls or active engagement can reveal the specific reasons for lurking. If patterns emerge, ways to activate these members become obvious.

Things to watch out for are:

  • Is the number of active participants growing, stalling or falling?
  • How is the onboarding experience structured? Is it enticing and easy to understand?
  • Are closed cliques forming who hog tasks and are unfriendly to newcomers?
  • Are chats flooded with messages and conversations becoming hard to follow?

Dunbar’s number is an important concept to keep in mind when thinking about engagement. Too much participation without structure can destroy any chance of meaningful conversations and relationship building, as anyone who’s been in a Telegram group of an airdrop can confirm.

If a DAO is growing and chats start to get flooded with messages, governance should think about opening channels with a specific focus in Discord, or even start Sub-DAOs that focus on a subset of the broader mission. This is an excellent problem to have, but it must be addressed. If conversations get swept away by a flood of other users posting, less assertive members leave, and some of those can be star contributors, if they can find a space to blossom.

#general, #off-topic and #introduction chats can usually not be broken down into smaller sub-chat. But pinned messages or Discord bots can make sure that new users quickly find their way to the subgroups most relevant to them. These can then be managed to remain small and comfortable places where meaningful conversations happen and work gets done.

DAO culture, mythology and the art of storytelling

When DAOs are small, and lines of communication between members are direct, maintaining a common purpose is easy. As more and more members join, interpretations of the organization’s purpose start to diverge, weakening the impact of the whole enterprise.

This is especially true for large, distributed organizations, with little formal onboarding and where requiring members to attend meetings is hard. Regular community calls are invaluable, but a call with 12,000 attendees means only a few questions will get answers, and only some opinions heard.

Without a strong sense of purpose and active community-building, DAOs tend to become chaotic at first, and boring a short while later, as members simply give up on trying to participate. Since DAOs seldom offer regular paychecks, community and friendship are key. Governance must ensure that this can easily happen. Buddy systems, newcomer events in small groups and access-gated channels are some ways that can help members get to know each other better and foster meaningful relationships. Ultimately it will be the relationships that bind members and provide the most value. The cohesion of a DAO is what governs its staying power in a competitive environment.

Gaming guilds are a wonderful example where a common purpose and a clearly defined mission are shared. A common passion means new members have common talking points that help with bonding.

DAO governance should be aware of:

  • The common purpose and story of the DAO are reiterated often enough and are dialled in.
  • The mission of the DAO is refined periodically and is well understood.
  • There is a common ground and common values that can be built upon.
  • A higher purpose and calling can be identified.

To call a common purpose or a shared story of a DAOs place in the world organizational mythology might seem hyperbole, but we think it fits the bill. Humdrum everyday stories do not have the necessary appeal and power to motivate members in the long term. The power of myth is precisely in its vagueness, which helps it adapt to a much wider variety of situations than rigorous prose.

Take Apple, for instance: by consistently telling their brand story over decades, users believe in Apple making the best products for successful creatives and being at the forefront of innovation. They are more likely to forgive occasional blunders, like faulty keyboards, because they believe that “Apple are the good guys”. This is what organizational mythology can do. A certain attitude and aura are associated with an organization, and this vague notion is then self-actualizing, as long as the core message is not directly contradicted by the actions of the organization.

Governance should define early on:

  • What common values the DAO should stand for.
  • If there is a founding story worth telling.
  • If there is a higher calling worth fighting for.
  • How this identity is maintained with storytelling and how it is refreshed and actualized.

DAO governance should look out if these values emerge from within the organization and capitalize on them. Community engagement should relentlessly reiterate the story that defines a DAOs purpose because stories bind human communities in a more sustainable way than rational objectives.

Conclusion

All organizations are networks. DAOs just allow people to join and leave faster. The power of DAOs is that the barrier to entry is extremely low. But so is the barrier to leaving.

Network effects are at the heart of growing and maintaining a vigorous distributed organization that ships great work. DAO governance needs to make sure they understand these dynamics and set appropriate incentives.

We want to finish this article with a quote by American author and entrepreneur Keith Ferrazzi

"The currency of real networking is not greed but generosity."

May generosity and a higher calling inform your decisions.

Arweave TX
rVFT9-GW9Dm9i6nRpNUcI3gjBNOVbpWhRpUFlf4b52A
Ethereum Address
0xaeE6b099f09142f95d0680C379c82C0A3D306a64
Content Digest
TiB-RCiZCAWj1WVcuDYHOjvbkCbJl8R7Bj3qmgp2dnc